Games with Models, 5/19/20 Data

COVID19-hospitalizationsandprojections-051920-featured

Unfortunately, today’s data saw an increase in hospitalized Rhode Islanders who tested positive for COVID-19 and a larger increase in the number of deaths.  However, it’s especially relevant today that the hospitalization numbers are actually two days behind the other numbers the state government is providing each day.

Thus, the increase actually happened from Saturday to Sunday.  and looking more closely the issue wasn’t more people admitted to the hospital so much as fewer discharged.  The number of people admitted to the hospital (22) was actually a little lower than the two-week average, but the number of people discharged (11) was well below the average.  Of course, the day before, the number discharged (40) was well above the average, so the two-day average might have been more relevant this weekend for some reason.

In any event, as the daily chart shows, the upshot was just to bump the hospitalization trend line back to where it had been the day before the big decrease.

COVID19-hospitalizationsandprojections-051920

 

(See here for my original methodology and here for a subsequent modification I made.)

Projections versus actuals (date of report).

  • Cases:
    • Projection for 5/19: 12,919
    • Actual for 5/19: 12,951
    • Projection for 5/20: 13,076
  • Hospitalizations:
    • Projection for 5/19: 224
    • Actual for 5/19: 247
    • Projection for 5/20: 236
  • Deaths:
    • Projection for 5/19: 511
    • Actual for 5/19: 532
    • Projection for 5/20: 537

Note that the big swing in deaths was actually spread over multiple days of revisions. The one-day projection had been for five more deaths, and there were 14, but not the 26 reported.